The effect of gelatin inclusion in high protein extruded pet food on kibble physical properties

The effect of gelatin inclusion in high protein extruded pet food on kibble physical properties

Accepted Manuscript Title: The effect of gelatin inclusion in high protein extruded pet food on kibble physical properties Authors: A.E. Manbeck, C.G...

281KB Sizes 35 Downloads 128 Views

Accepted Manuscript Title: The effect of gelatin inclusion in high protein extruded pet food on kibble physical properties Authors: A.E. Manbeck, C.G. Aldrich, S. Alavi, T. Zhou, R.A. Donadelli PII: DOI: Reference:

S0377-8401(17)30195-5 http://dx.doi.org/doi:10.1016/j.anifeedsci.2017.08.010 ANIFEE 13841

To appear in:

Animal

Received date: Revised date: Accepted date:

10-2-2017 15-6-2017 7-8-2017

Feed

Science

and

Technology

Please cite this article as: Manbeck, A.E., Aldrich, C.G., Alavi, S., Zhou, T., Donadelli, R.A., The effect of gelatin inclusion in high protein extruded pet food on kibble physical properties.Animal Feed Science and Technology http://dx.doi.org/10.1016/j.anifeedsci.2017.08.010 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

The Effect of Gelatin Inclusion in High Protein Extruded Pet Food on Kibble Physical Properties

A.E. Manbeck, C.G. Aldrich, S. Alavi, T. Zhou, and R.A. Donadelli

Department of Grain Science Kansas State University 201 Shellenberger Hall Manhattan, KS 66506 [email protected] [email protected] [email protected] [email protected] [email protected]

Corresponding Author: Greg Aldrich, [email protected]

Highlights: 

Low-bloom gelatin may act as a nutritional binder in dry extruded pet food.



Initial experiments showed increased kibble durability with gelatin inclusion.



Further experimentation showed increased kibble expansion with gelatin inclusion.



Gelatin may have a two-fold effect on kibble depending on water inclusion.

Abstract Pet food is a $23 billion industry that continues to grow. Owners continue to humanize their pets and their dietary needs, thus the pet food industry tends to mirror human dietary trends. Currently, pet food is trending towards higher levels of protein, thus lower levels of starch. Decreasing starch, one of the main structure forming ingredients in extruded foods, creates issues of lower rates of expansion and decreased kibble durability. Consumers tend to dislike ingredients that do not serve a nutritional purpose; therefore, gelatin may be a plausible binding 1

ingredient for high protein pet foods. Gelatin is a pure protein derived from collagen and is sold as a dry, odorless, tasteless powder. High-bloom gelatins find numerous uses in human food as a stabilizer, foaming agent, and capsule base among other uses. Low-bloom gelatin may find a value-added opportunity as a nutritional binder in pet food. Four extrusion experiments were performed to test this hypothesis. Experiment 1 compared gelatin at 0, 50, 100, and 150 g/kg inclusion and 150 g/kg gelatin at 3 different extruder screw speeds. Results showed a decrease in expansion but an increase in hardness and pellet durability index (PDI); however, there may have been inadequate preconditioning. It was unclear whether the decrease in expansion or presence of gelatin improved product durability. Experiment 2 analyzed two levels of gelatin, 0 and 100 g/kg, under two extruder screw speeds, 300 rpm and 500 rpm, and two hydration ratios, 170 and 280 g/kg. In this experiment, there were no differences in density, expansion, hardness, or PDI. This indicated preconditioning was more ideal and may indicate gelatin does not decrease product expansion. Experiment 3 analyzed two levels of gelatin, 0 and 100 g/kg, at two target densities, low and high. Results indicated that gelatin created a more expanded product when processed under similar conditions as a control formula. Experiment 4 analyzed different strengths of gelatin to determine if the low-bloom gelatin experiments were repeatable with more conventional strength gelatins. Treatments were a control with no gelatin, and a 100, 175, and 250 bloom gelatins. Results showed increased gelatin strength increased product expansion, likely through a foaming effect. However, durability declined with mid- and high-bloom gelatins; thus, low-bloom gelatin may be the most promising to improve product characteristics and preserve durability. Based on these experiments, low-bloom gelatin may find use as a nutritional binder in high protein pet foods. Keywords: Pet Food, Gelatin, Extrusion 1.0 Introduction Pet food is a rapidly growing industry. The industry grew from $22.26 billion in 2014 to just over $23 billion estimated in 2015. In the current economy, few industries continue to see steady growth like the pet food industry. Pet owners are spending an average of $269 per year on dog food and $246 per year on cats (APPA, 2015). Humanization of pets spurs trends that drive the growth of the industry. Human food trends are currently driving popular diets that promote increased protein consumption and tend to 2

eschew traditional cereal grain sources of carbohydrates (Mileen et al., 2015). Naturally, these trends have been paralleled in the pet food industry due to the close connections between owners and their pets. The issue with increasing the protein content of pet food is apparent during processing. Increasing the protein content inherently decreases the starch content. When starch sources or plant-based proteins are replaced by animal proteins, products become less expanded, denser, and more fragile. For example, Zhu et al. (2010) reported that high levels of soy protein caused low levels of expansion due to starch-protein interactions where the proteins interfered with the continuous matrix of the product. In pet food, this is an issue in multiple ways from processing through to consumer use. Typically, extruded products do not use binders or ingredients strictly to improve the cohesion of the internal matrix; in part because starches are usually included in high enough proportions to provide significant structural enhancement. With the push towards more natural pet foods, consumers are leery about ingredients that have no nutritional value. If they are missing because of market drivers, then we need an alternative like low-bloom gelatin to fill the void. While low bloom gelatin has yet to be used in extruded foods, there may be an opportunity for its incorporation as a binding ingredient in higher protein pet food. As a nearly pure protein, gelatin provides nutritional value and it may have a functional use within a food matrix to improve structural durability. Therefore, our objectives were to determine the effect of low bloom gelatin in a moderate protein extruded pet food diet on processing parameters and physical properties of the kibble. 2.0 Materials and Methods Two initial experiments were performed to determine the changes in kibble physical properties when gelatin was used at different inclusion rates in a dry extruded pet food. Experiment 1 (Table 1) tested 4 levels of gelatin – 0 (OG), 50 g/kg (5G), 100 g/kg (10G), and 150 g/kg (15G) – and three different extruder screw speeds on the 15G diet – a high screw speed of 500 rpm (15GH), a moderate screw speed of 400 rpm (15GM), and a low screw speed of 300 rpm (15GL). Experiment 2 (Table 2) tested 2 levels of gelatin – 0 and 100 g/kg - 2 screw speeds – low speed at 300 rpm and high speed at 500 rpm – and two different hydration ratios – 170 g/kg and 280 g/kg. Hydration ratio is expressed as the proportion of total water added as water 3

into the extruder. The 170 g/kg hydration ratio was achieved with 11 kg/hr steam in the preconditioner, 14 kg/hr water in the preconditioner, and 5 kg/hr water in the extruder. The 280 g/kg hydration ratio was achieved with 9 kg/hr steam in the preconditioner, and 11 kg/hr water in the preconditioner, and 8 kg/hr water in the extruder. Hydration ratio can also be described as 2.7% (170 g/kg) or 4.4% (280 g/kg) of total water directed to the extruder. Two additional experiments were performed to determine the relationship between gelatin level and product density, and to determine the effect of increasing gelatin bloom strength on kibble properties. Experiment 3 (Table 3) compared 0 gelatin (0G) with 100 g/kg gelatin (10G) at a low target density (LD) and a high target density (HD). Experiment 4 (Table 4) compared a control diet with 0 g/kg gelatin (OG) with 3 strengths of gelatin – low bloom (100G), mid bloom (175G) and high bloom (250G). These diets were processed once through a circle die (C) and once through a triangle die (T). 2.0.1 Diet Preparation Diets for all experiments were formulated according to the American Association of Feed Control Officials Publications dog food nutrient profiles for all life stages (AAFCO, 2015). Four diets were formulated with four different levels of a gelatin for use in all four experiments (Table 5): 0G was the control formula and contained no gelatin, 5G contained 50 g/kg gelatin, 10G contained 100 g/kg gelatin, and 15G contained 150 g/kg gelatin. Gelatin was added at the expense of chicken by-product meal to maintain consistent levels of starch in all diets. Select micro-ingredients were added to 5G, 10G, and 15G to keep the entire diet isonutritional. Further, in Experiment 1, only a small amount of chicken fat was also added to 5G, 10G, and 15G (Table 6). Experiments 1, 2, and 3 utilized low-bloom gelatin, and experiment 4 utilized low-, mid-, and high-bloom gelatins. The low-bloom gelatin used was a 50-bloom pork bone gelatin (Pro-Bind Plus 100, Sonac, Darling Ingredients International, Irving, TX). Mid-bloom and high bloom gelatins were 171-bloom pig skin and 250-bloom pig skin (Pig Skin 175 and Pig Skin 250, Rousselot, Darling Ingredients International, Irving, TX). Bloom strengths were gathered from the Certificate of Analysis that accompanied each batch of gelatin. All other ingredients were purchased from a local mill (Lortscher Animal Nutrition, Bern, KS). Brewers’ rice, corn, wheat, and beet pulp were ground to pass through a 0.50 mm screen (Fitzmill hammermill, Continental Agra 4

Equipment Inc., Newton, KS). These ingredients, along with chicken by-product meal and gelatin, were weighed on a large digital scale with 0.01 kg sensitivity and then mixed in a double ribbon mixer for 5 minutes. Potassium chloride, salt, choline chloride, DL methionine, dry natural antioxidant, trace mineral premix, vitamin premix, calcium carbonate, taurine, monosodium phosphate, and L-tryptophan were weighed according to the formula using a laboratory scale with 0.001 g sensitivity and combined in a small container and added to the double ribbon mixer and mixed for an additional 3 minutes. In experiment 1, chicken fat was added to 5G, 10G, and 15G slowly into the ration while the mixer was running. Once the fat was added, the diet was mixed for an additional 2 minutes. In all other experiments, chicken fat was excluded from the formulation as it would typically be added topically to the kibbles following processing. The addition of gelatin, and subsequent removal of chicken by-product meal, caused a decrease in calculated fat content of the 5G, 10G, and 15G diets. This would have been remediated with an increase in chicken fat inclusion; however, the chicken fat would have been topically applied and not introduced during extrusion processing. Products were not used in feeding studies, therefore were not coated with any external fat. Following production diet treatments samples were stored in poly-lined paper bags (25 kg).

2.0.2 Extrusion Conditions Applicable to all Experiments Diets for all experiments were processed sequentially through a differential diameter cylinder preconditioner, single screw pilot-scale extruder, and double pass dryer (DDC Preconditioner, X-20 Extruder, and 4800 Series Dryer, Wenger Manufacturing, Sabetha, KS). The screw configuration used in both experiments consisted of the following: extruder inlet, single flight uncut, small steam lock, single flight uncut, small steam lock, single flight uncut, medium steam lock, double flight uncut, large steam lock, and double flight uncut cone. The dryer was set at 104°C with 5 minutes for each of the three passes. In-Barrel Moisture (IBM) was calculated using the following equation: IBM (%) = [(Mm*Mt + Ps + Pw + Es + Ew) / (Mt + Ps + Pw + Es + Ew)] *100 Where: Mm = material moisture content Mt = material throughput 5

Ps = preconditioner steam Pw = preconditioner water Es = extruder steam Ew = extruder water Specific Mechanical Energy (SME) was calculated using the following equation: SME (kJ/kg) = [(%T - %T0) * (Pr) * (N/Nr)] / m Where: %T = motor load or percent torque at full load %T0 = motor load or percent torque at no load Pr = rated motor power N = measured screw speed Nr = rated screw speed m = throughput 2.1.0 Experiment 1 Processing Conditions Treatments were labeled according to gelatin inclusion and extruder screw speed: 0 g/kg gelatin at 400 rpm (0GM), 50 g/kg gelatin at 400 rpm (50GM), 100 g/kg gelatin at 400 rpm (100GM), 150 g/kg gelatin at 300 rpm (150GL), 150 g/kg gelatin at 400 rpm (150GM), and 150 g/kg gelatin at 500 rpm (150GH). The following were the processing conditions used in experiment 1: dry feed rate – 180.0 kg/hr, feeder screw speed – 18.0 rpm, preconditioner cylinder speed – 400 rpm, preconditioner water – 18.0kg/hr, preconditioner steam – 15.0 kg/hr (8.2%), preconditioner discharge temperature – 96°C, extruder water – 6.0kg/hr (3.3%), temperature zone 1 - 50°C, temperature zone 2 - 70°C, temperature zone 3 - 90°C, and knife speed – 1360 rpm. Extruder screw speed varied according to the treatment, with 4 treatments being processed at a moderate screw speed of 400 rpm, one treatment being processed at a low screw speed of 300 rpm, and one treatment being processed at a high screw speed of 500 rpm. Resulting calculated in-barrel moisture was 23.8%. Resulting calculated SME for each treatment was as follows: 0G – 138.5 kJ/kg, 5G – 110.3 kJ/kg, 10G – 111.1 kJ/kg, 15GH – 115.7 kJ/kg, 15GM – 97.5 kJ/kg, 15GL – 88.6 kJ/kg. The extruder was fit with a single 3.19 mm diameter circle die, resulting in 7.99 mm2 open area, and a hard knife assembly. 6

2.2.0 Experiment 2 Processing Conditions Treatments were labeled according to gelatin inclusion, extruder screw speed, and hydration ratio: 0 g/kg gelatin at 300 rpm and 170 g/kg hydration ratio (0GL17), 0 g/kg gelatin at 500 rpm and 170 g/kg hydration ratio (0GH17), 0 g/kg at 500 rpm and 280 g/kg hydration ratio (0GH28), 100 g/kg gelatin at 300 rpm and 170 g/kg hydration ratio (10GL17), 100 g/kg gelatin at 500 rpm and 170 g/kg hydration ratio (10H17), and 100 g/kg gelatin at 500 rpm and 280 g/kg hydration ratio (10GH28). The following processing conditions were used in experiment 2: dry feed rate – 180.0 kg/hr, feeder screw speed – 18.0 rpm, preconditioner cylinder speed – 400 rpm, temperature zone 1 - 50°C, temperature zone 2 - 70°C, temperature zone 3 - 90°C, and knife speed – 1428 rpm. Extruder screw speed, preconditioner steam, preconditioner water, and extruder water varied according to treatment. Extruder screw speed was 300 rpm for two treatments and 500 rpm for four treatments. Preconditioner steam, preconditioner water, and extruder water was adjusted according to treatment to achieve the 170 g/kg and 280 g/kg hydration ratios. The 170 g/kg hydration ratio used for 4 treatments was achieved by injecting 11 kg/hr of preconditioner steam, 14 kg/hr of preconditioner water, and 5 kg/hr of extruder water. This resulted in an in-barrel moisture content of 21.1% with 17% (or 170 g/kg) of total moisture being added as water into the extruder. This treatment also resulted in a preconditioner discharge temperature of 85°C. The 280 g/kg hydration ratio used for 2 treatments was achieved by injecting 9 kg/hr of preconditioner steam, 11 kg/hr of preconditioner water, and 8 kg/hr of extruder water. This resulted in an in-barrel moisture content of 20.4% and 28% (or 280 g/kg) of total moisture being added as water into the extruder. This treatment also resulted in a preconditioner discharge temperature of 75°C. The resulting calculated SME for each treatment was as follows: 0GL17 – 199.6 kJ/kg, 0GH17 – 183.5 kJ/kg, 0GH28 – 136.5 kJ/kg, 10GL17 – 199.6 kJ/kg, 10GH17 – 146.8 kJ/kg, and 10GH28 – 123.3 kJ/kg. The extruder was fit with a single 3.19 mm diameter circle die, resulting in a die open area of 7.99 mm2, and a hard knife assembly.

7

2.3.0 Experiment 3 Processing Conditions Treatments were labeled according to gelatin inclusion and target product density: 0 g/kg gelatin at high density (OGHD), 0 g/kg gelatin at low density (OGLD), 100 g/kg gelatin at high density (10GHD) and 100 g/kg at low density (10GLD). The following were the constant processing conditions used: dry feed rate – 180.0 kg/hr, feeder screw speed – 18.0 rpm, preconditioner steam – 16.0 kg/hr, preconditioner water – 14.0 kg/hr, preconditioner discharge temperature – 95°C, extruder water – 11.0 kg/hr, temperature zone 1 – 50°C, temperature zone 2 – 70°C, temperature zone 3 – 90°C, and knife speed – 1793 rpm. Two different product densities, low density (LD) and high density (HD), were achieved by varying the extruder screw speed and the opening of a throttle valve fitted between the end of the extruder and the die assembly. The OGHD was achieved with an extruder screw speed of 500 rpm and the throttle valve turned 5½ times resulting in an average die pressure of 550 psig (3.792 MPa). The OGLD was achieved with an extruder screw speed of 550 rpm and the throttle valve turned 7 times resulting in an average die pressure of 500 psig (3.447 MPa). 10GHD was achieved using an extruder screw speed of 500 rpm and the throttle valve turned 4 times resulting in an average die pressure of 380 psig (2.620 MPa). 10GLD was achieved using an extruder screw speed of 550 rpm and the throttle valve turned 5½ times resulting in an average die pressure of 390 psig (2.689 MPa). The resulting calculated in-barrel moisture was 25.1%. The resulting calculated SME for each treatment was as follows: 0GHD – 110.1 kJ/kg, 0GLD – 153.4 kJ/kg, 10GHD – 168.8 kJ/kg, 10GLD – 161.5 kJ/kg. The extruder was fit with a single 4.6 mm diameter circle die, resulting in a die open area of 16.62 mm2, and a hard knife. 2.4.0 Experiment 4 Processing Conditions Treatments were labeled according to gelatin type and die shape: no gelatin by circle die (0GC), no gelatin by triangle die (0GT), Pro-Bind 100 gelatin by circle die (100GC), Pro-Bind 100 gelatin by triangle die (100GT), Pig Skin 175 gelatin by circle die (175GC), Pig Skin 175 gelatin by triangle die (175GT), Pig Skin 250 gelatin by circle die (250GC), and Pig Skin 250 gelatin by triangle die (250GT).

8

The following are the processing conditions used across all treatments: dry feed rate – 165.0 kg/hr, feeder screw speed – 16.2 rpm, preconditioner cylinder speed – 400 rpm, preconditioner steam – 16.0 kg/hr, preconditioner water – 15.0 kg/hr, preconditioner discharge temperature – 96°C, extruder screw speed – 545 rpm, extruder water – 16.0 kg/hr, temperature zone 1 – 50°C, temperature zone 2 – 70°C, and temperature zone 3 – 90°C. The throttle valve was turned down 5 turns for all treatments, and the resulting die pressure ranged from 440-500 psi (3.034 – 3.447 MPa). The resulting in-barrel moisture was 28.4%. The resulting calculated SME for each treatment was as follows: 0GC – 148.4 kJ/kg, 0GT – 183.3 kJ/kg, 100GC – 157.1 kJ/kg, 100GT – 148.4 kJ/kg, 175GC – 187.6 kJ/kg, 175GT – 165.8 kJ/kg, 250GC – 261.8 kJ/kg, and 250GT – 200.7 kJ/kg. All 4 diets were first processed through a single 4.5 mm diameter circle die, with a die open area of 15.9 mm2, in a randomized order. The extruder was then shut down and the die was replaced with a single 5.3 mm base by 4.9 mm height triangle, with a die open area of 12.99 mm2, and all 4 diets were processed a second time in a randomized order. Knife speed for all circle-shaped treatments was 2,046 rpm and knife speed for all triangle-shaped treatments was 2,038 rpm. 2.0.3 Product Analysis All samples were stored in a freezer (-20°C) after drying and prior to testing to prevent moisture loss. Samples were warmed to room temperature for at least 2 hours prior to analysis. Products in Experiment 1 were analyzed for post-extrusion and post-drying moisture, postextrusion and post-drying expansion ratio, post-extrusion and post-drying specific length, postdrying piece density, post-drying hardness, and post-drying pellet durability index (PDI). Products in Experiment 2 were analyzed for post-extrusion and post-drying moisture, postextrusion and post-drying expansion ratio, and post-drying specific length. Products in Experiments 3 and 4 were analyzed for bulk and piece density, radial expansion, specific length, hardness, and pellet durability index (PDI). Moisture was analyzed using the Association of Official Analytical Chemists (AOAC) method 934.01 (AOAC, 2010). Extrudates were first ground, and then 2 g of sample was dried at

9

95-100°C for 4 hours and the loss in weight was reported as moisture. Moisture was expressed in g/kg. Expansion ratio represents the ratio of the size of the product relative to the size of the die to capture the degree of expansion that occurs in the radial or axial direction. Calipers were used to measure the diameter of the sample in triplicate. Expansion ratio (ER) was then calculated using the following formula: ER = De2 / Dd2 Where De2 is the average diameter of the extrudate squared and Dd2 is the average diameter of the die squared. A higher value for expansion ratio corresponds to a more expanded, less dense product. Expansion ratio was expressed in mm2/mm2. Specific length represents expansion in the longitudinal direction and is a ratio of the length of the sample to the weight of the sample. Calipers were used to measure the length of the samples, and a laboratory scale with 0.001 g sensitivity was used to weigh the samples. Specific length was expressed in cm/g. Bulk density represents the ratio of the weight of the product to the volume of the product on a small bulk sample basis. An empty 1 liter cup was tared on a scale with 0.01g sensitivity. Product was collected in the 1 liter cup without shaking or packing the sample. The sample was then weighed. Bulk density is expressed in g/L. Piece density represents the ratio of the weight of the product to the volume of the product on a single piece basis. Calipers were used to measure the diameter and height or length of the product. A lab scale with 0.001g sensitivity was used to weigh the samples. Piece density (PD) was then calculated using the following formula: PD = Weight / (πr2h) Where πr2h is the volume of a cylinder and r is the radius and h is the height, or in the case of extruded product the length. Piece density is expressed in g/cm3. Hardness was determined using a texture analyzer (TA-XT2 with Exponent 32 software, Stable Micro Systems, Godalming, UK) fitted with a 38.1 mm diameter, flat, plastic compression probe (Texture Technologies, Hamilton, MA). A compression test with a pre-test and test speed of 1.0 mm/s, post-test speed of 2.0 mm/s, and 600 g/kg strain was used on 20 pieces of sample. Hardness is expressed in kg of force required to compress the sample.

10

Pellet durability index (PDI) was analyzed using a pellet durability tester (Holmen NHP 100, Tekpro, Norfolk, UK) with 100 g of sample for 90 seconds. This instrument holds the sample in a perforated metal basket and forces air through the sample, causing the sample to impact the sides of the basket. The pneumatic force also carries any fines away from the remaining sample. At the end of the test, the remaining intact sample is weighed, and PDI is expressed as a percent (%) based on the amount of intact kibble relative to the starting sample weight. 2.0.4 Statistical Analysis Experiment 1 was an initial assessment of extruder settings and gelatin level and was conducted as a completely randomized design and processed in the order of 0, 50, 100, and 150 g/kg at 400 rpm followed by 150 g/kg at 500 rpm, and then 150 g/kg at 300 rpm. Results were analyzed using the MEANS procedure of SAS (SAS Institute, Cary, NC) and reported as simple means. Experiment 2 was a further exploration of necessary process settings and was arranged in an incomplete factorial arrangement of treatments (Byar, 1993) evaluating two levels of gelatin (0 and 100 g/kg), two extruder screw speeds (300 and 500 rpm), and two hydration ratios (170 and 280 g/kg). Sample time (3 per treatment at 15 minute intervals throughout the 30 minute run following achievement of steady state) was considered to be the experimental unit. Results were analyzed using the GLIMMIX procedure of SAS (SAS Institute, Cary, NC) separately for each main effect, but due to inadequate replication within treatments no interactions were evaluated. Main effect means were separated by least square means and considered significant at an alpha of 5%. Experiment 3 was arranged in a 2x2 factorial design structure with two levels of gelatin (0 and 100 g/kg) and two target densities (High and Low). Sample time (3 per treatment at 15 minute intervals throughout the 30 minute run following achievement of steady state) was considered to be the experimental unit. Results were analyzed using the GLM procedure of SAS (SAS Institute, Cary, NC). Main effect means and the interaction of the two were separated by least square means and considered significant at an alpha of 5%. Experiment 4 was arranged in a 4x2 factorial arrangement of treatments with four levels of gelatin bloom strength (0 gelatin, and 100, 175, and 250 bloom) and two die shapes to create 11

replication. Results were analyzed using the GLM procedure of SAS (SAS Institute, Cary, NC). Main effect means for die shape were not different for each variable and were pooled thereafter. Data for the pooled main effect of bloom strength was analyzed by orthogonal single degree of freedom contrasts for control vs gelatin containing formulas, and linear, quadratic, and cubic effects of gelatin inclusion. While reported, the cubic contrasts were not discussed due to limited number of treatment ranges. 3.0 Results 3.1 Experiment 1 Moisture and piece density held relatively constant as gelatin level and extruder screw speeds changed (Table 7). Post-extrusion and post-drying expansion ratios declined numerically with increased gelatin level. Results of post-extrusion and post-drying expansion ratios were inconsistent as extruder screw speed increased. Post-extrusion and post-drying specific length did not seem to be influenced much by gelatin strength. However, specific length post-extrusion and post-drying were somewhat greater for the lower screw speed than the medium or high. Hardness and PDI increased somewhat with the first increment of gelatin, but additional amounts did not appear to have a clear influence. The hardness, but not PDI, appeared to be affected by increasing screw speed 3.2 Experiment 2 In the second experiment the post-extrusion and post-drying moisture (n=12 for postextrusion and post-drying analysis) were not affected (P>0.05) by gelatin level, screw speed, or hydration ratio (Table 8). Likewise, post-extrusion and post-drying expansion ratio (n=30 for post-extrusion and post-drying analysis) was not affected (P>0.05) by gelatin level, screw speed, or hydration ratio (Table 9). Specific length (n=30) was not affected (P>0.05) by gelatin level or screw speed, but was affected (P<0.05) by hydration ratio (Table 10). In this case, specific length (n=30) decreased (P<0.05) as the proportion of water in the extruder increased (4.12 cm/g for 170 g/kg vs. 3.76 cm/g for 280 g/kg). 3.3 Experiment 3 The bulk density (n=20) and piece density (n=40) were lower (P<0.05) for 10G (Table 11). Bulk (n=20) and piece density (n=40) decreased (351.83 g/L vs. 280.67 g/L and 0.52 g/cm3 12

vs. 0.39 g/cm3, respectively). Radial expansion (n=40) increased (P<0.05) from 2.86 mm2/mm2 for 0G to 3.56 mm2/mm2 for 10G and the specific length (n=40) also increased from 4.12 cm/g for 0G to 4.47 cm/g for 10G. Though hardness (n=40) was unaffected, PDI (n=20) decreased (P<0.05) in the gelatin containing diet (77.57 % vs. 52.25 % for 0G and 10G, respectively). High target density corresponded to higher bulk density (n=20) (P<0.05; 345.83 vs. 266.67 g/L for high density and low density, respectively). Piece density (n=40) had a similar relationship to target (Piece density target was 0.460 g/cm3 vs. 0.441 g/cm3 for high and low density, respectively). There was no effect (P>0.05) on radial expansion (n=40), specific length (n=40), or hardness (n=40) (Table 12). However, the PDI (n=20) was lower for the low target density (72.17 % vs. 57.65 % for high and low target density, respectively). 3.4 Experiment 4 Bulk density was higher (P<0.05) for the 0 vs gelatin containing treatments (346.4 vs average 257.0 g/L; Table 13). As well, there was a linear (P<0.05) and quadratic (P<0.05) decrease in bulk density as gelatin bloom strength increased. Piece density, cross-sectional expansion, hardness, and PDI followed a similar relationship among treatments. The specific length was shorter for 0 vs the gelatin containing treatments (4.27 vs average 4.50 cm/g), but did differ in a linear fashion as bloom strength increased. 4.0 Discussion The issue with increasing the protein content of pet food can be apparent during processing; wherein, increasing the protein content directly decreases the starch content. This has an impact on the kibble physical stability or integrity. In the following experiments, the diet formulation used was of moderate protein content and contained multiple starch sources. This formulation was utilized to represent an average canine diet and obtain baseline results as to the effect of gelatin inclusion. Further experiments should focus on base formulations higher in protein and lower in starch. 4.1 Experiment 1 Experiment 1 was a preliminary evaluation intended to provide a first look at how increasing low-bloom gelatin would influence process and piece characteristics. The extruder screw speed for the150 g/kg gelatin diet was also changed to determine the impact it would have at the extreme gelatin addition level where the most likely detectable difference might occur. As 13

samples were pulled from the preconditioner it was observed that increased gelatin led to increased particle agglomeration. Gelatin is a very hydrophilic substance which has been reported to absorb 5-10 times its volume of water (Baziwane and He, 2003). This was apparent during the preconditioning phase as the gelatin portions of the mixtures appeared to absorb the majority of water injected. The water absorption and starch gelatinization were not quantified during this experiment, but we speculate that it may have been responsible for the decline in product expansion and increased hardness. Taking this one step farther, if the preconditioning moisture was primarily absorbed by the gelatin then the starches in the ration were likely not hydrated sufficiently for proper expansion (Dogan et al., 2010). Improper hydration during preconditioning impacts overall starch gelatinization and leads to reduced product expansion and textural changes (Chuang and Yeh, 2004). Further, there were no differences in moisture content for both post-extrusion and post-drying, across all treatments. The true reason behind the inconsistent expansion ratio with increases in gelatin and (or) screw speed is not fully understood. It may relate to the inadequate hydration of the raw material observed from the preconditioner mash samples which is known to directly affect radial expansion of starches (Chuang and Yeh, 2004). Furthermore, increasing extruder screw speed creates more mechanical energy and more pressure behind the die which usually results in a higher rate of expansion due to the added steam flash-off (Dogan et al., 2010). Conversely, lower screw speeds may increase residence time within the extruder, but don’t create as much added pressure to increase expansion. In this experiment, expansion seemed to increase at both a lower screw speed of 300 rpm and a higher screw speed of 500 rpm relative to the moderate screw speed of 400 rpm. The higher expansion at the 500 rpm screw speed was expected. It was hypothesized that that slowing the extruder screw speed to 300 rpm may have increased the residence time and in turn allowed for more contact between the preconditioned material and the added water injected in the extruder. It was also noted during the experiment that preconditioning could have been incomplete due to the presence of gelatin and its more rapid absorption of the moisture relative to the starches. Therefore the increase in residence time in the extruder may have acted as an additional conditioning step and aided in the gelatinization of the starches; thus, increasing the final product expansion. Based on these results, gelatin may improve the expansion of pet food, but more

14

exploration is needed to clarify the ideal preconditioner and extruder settings to optimize processing with formulations including gelatin. The relationship between gelatin level and specific length suggests that gelatin may have a plasticizing effect which is more prominent with a higher level of gelatin in the formula (Wulansari et al, 1999). This added lubrication of the extruder barrel might have decreased the amount of mechanical energy incorporated into the product. While the length of the product could be adjusted by altering the knife speed, the increased rate of velocity of product through the die might also contribute to the low radial expansion of the 150 g/kg gelatin formula. Extruder screw speed also affected specific length. Typically, an increase in extruder screw speed would be expected to increase specific length as the added mechanical energy increases pressure behind the die and thus increases product velocity. However, the presence of gelatin may create a plasticizing effect that is exaggerated at a lower screw speed. Furthermore, the lower screw speed may have allowed for added hydration of the gelatin. The longer residence time and added period of hydration may have allowed the gelatin to form more of a gel that coated the barrel for the “plasticizing-like” effect. The increase in hardness and PDI with the addition of gelatin indicated that gelatin could improve the durability of high protein kibble. However, a decrease in product expansion inherently makes a product harder and more durable because the internal air cells decrease in size and the cell walls thicken which naturally strengthens the product matrix. Further experiments to clarify if the increase in durability was caused by the inclusion of gelatin or if it was simply a factor of the decrease in expansion were conducted. 4.2 Experiment 2 Taking this forward, experiment 2 was organized to evaluate two levels of gelatin (0 and 100 g/kg), processed under two screw speeds (300 and 500 rpm) and two hydration ratios (170 and 280 g/kg of total moisture added at the preconditioner; Tables 9-11). The post-extrusion and post-drying moisture levels were unaffected by treatment. It should be noted that a change in moisture might be expected with the change in hydration ratio, but total moisture added, and subsequent in-barrel moisture content, was not altered. Rather, the distribution of water and steam inclusion between the preconditioner and extruder was shifted to alter treatment conditions. Even yet, it was surprising that there were no changes in either expansion ratio or specific length despite formulation and processing changes. However, all products had a rather 15

high expansion ratio. In the previous experiment, the inclusion of gelatin led to a decrease in expansion; whereas, in this experiment, the inclusion of gelatin had no effect. This may indicate that the overall hydration of the raw material was adequate in this experiment, but preconditioning was still inadequate in the presence of gelatin. Screw speed also had no effect on post-extrusion or post-drying expansion ratio. In the previous experiment, the slower screw speed increased product expansion likely because of the increase in conditioning perhaps due to the longer residence time. In this experiment, the product expansion similarity despite screw speed changes also indicates a more adequate preconditioning process. Specific length previously increased with the addition of gelatin indicating it may have imparted a plasticizing effect. In this experiment, specific length was not affected by gelatin level or screw speed (Table 10). There was a slight difference in specific length due to changes in hydration ratio. Typically, increasing the amount of water in the extruder, as was done with the 280 g/kg hydration ratio, would have increased specific length because water can act as a plasticizer. However, the reverse was observed in this experiment and again may be reflective of the different preconditioning requirements for gelatin. These two experiments provided a baseline of information about the use of low-bloom gelatin in a pet food, but demonstrated that more clarity was needed to further determine the relationship between gelatin inclusion and extruder processing conditions. Clarification was specifically needed regarding the relationship between gelatin inclusion and final product density. Based on experiments 1 and 2, the effect of gelatin on density may be dependent on the target density range. Experiment 1 demonstrated that gelatin may improve product durability in a low density range, but experiment 2 demonstrated that gelatin may improve expansion in a high density range. 4.3 Experiment 3 Experiment 3 compared 0 g/kg and 100 g/kg gelatin inclusion under similar processing conditions to achieve two distinctly different densities and provide more insight into the effect of gelatin on final product characteristics. First, it was noted during extrusion that a higher inclusion of water, particularly in the preconditioning step, resulted in what appeared to be a higher degree of cook as the kibble had a darker, more intertwined center. This greater addition of water may have influenced some of the 16

following results. Additional water likely allowed the gelatin to fully hydrate in the preconditioner while still hydrating starches and begin the gelatinization process as well. The inclusion of gelatin affected bulk and piece density, expansion both radially and longitudinally, and PDI, but surprisingly not hardness (Table 11). Bulk density and piece density decreased with the inclusion of gelatin. As was expected with this trend in density, radial expansion and specific length both increased in the presence of gelatin. The more complete hydration of the both the gelatin and the starches likely allowed for a more complete gelatinization thus greater rate of expansion with the 10G formula. These decreases in density and subsequent increases in expansion support a hypothesis of a protein foaming effect when gelatin was added (Miquelim et al., 2009; Muller-Fischer and Windhab, 2004; Schonauer et al., 2004). Gelatin, and some grain-based proteins, in the presence of a plasticizer have been shown to create a stable foam, particularly when carbon dioxide or nitrogen is introduced via a gas foaming process (Salerno et al., 2007). While water and steam were the only plasticizers (Guy, 2001) used in this experiment and there was no introduction of gases, it is possible the steam flash-off at the die may have had a similar effect to gas foaming. Typically, more expanded products are less durable. In this experiment, a decrease in PDI was seen, but interestingly there was no change in hardness. The higher rate of expansion may explain the decrease in PDI, however there may still be some effect of matrix enhancement due to gelatin inclusion that preserved hardness values. This hypothesis is somewhat reaffirmed when results were analyzed for the effects of high versus low density targets; wherein, low density led to a reduction in PDI but no change to hardness (Table 12). The decrease in PDI may be a result of the added expansion achieved with the addition of gelatin. This added expansion may be great enough that PDI cannot be preserved, while hardness is still somewhat aided by the binding ability of gelatin. The hypothesis established after experiment 2 regarding the possible dual effect of gelatin at different density ranges may be further supported by the results of experiment 3 in which the effect of gelatin appeared to be dependent on both formulation and processing conditions. 4.4 Experiment 4 The fourth and final experiment was conducted to determine if the results observed in the first three experiments using low-bloom gelatin would be the same using mid- and high-bloom gelatin. Stronger gelatins currently find more use in the food industry. 17

Two die shapes were used to add replication to the study. Some differences occurred; however it is likely that this was a function of slight differences in die open area and not a true treatment effect. Therefore, data for the die shape were pooled and main effect means for gelatin strength were reported (Table 13). As gelatin strength increased, density (bulk density and piece density) decreased and expansion increased (cross-sectionally and longitudinally). These changes in density, and expansion, point to the potential of a protein foaming effect. For example, the high-bloom gelatin, like the 250-bloom used in this experiment, is commonly used in confectionary applications such as marshmallows. Similarly when gelatin is combined with heat and air it creates a stiff foam that remains rigid as it cools (Baziwane and He, 2003). This effect might explain why the high-bloom gelatin created the most expansion and the expansion decreased as the strength of the gelatin declines. The increase in internal air cells and thinning of cell walls can create a weaker product. This relationship was seen when PDI was measured. However, hardness had a more interesting relationship, as it was strongest for the 100 bloom formula rather than the 0 bloom control. This indicated a stronger potential for binding capability when low-bloom gelatin was used. Overall, it appeared that the use of gelatin in pet food increased product expansion and this improved with the gelatin strength. Low-bloom gelatin showed the most promise for preserving or even enhancing product durability simultaneous to an increase in expansion. However, it should be noted that the rate of expansion for mid- and high-bloom gelatins was much higher than the average expansion ratio for pet food, particularly pet food with higher protein content. Therefore it may be possible to achieve the same durability preservation with mid- and high-bloom gelatins if used in lesser concentrations. 5.0 Conclusions Overall, gelatin increased the expansion of moderate protein pet foods and improved durability in most experiments, with sufficient preconditioner water addition, but may have different effects in depending on target density. Based on these experiments, low-bloom gelatin increased expansion while preserving the durability of the kibble. Gelatin may be a potential solution for durability and expansion issues experienced when producing higher protein pet foods. However, further experimentation is needed to evaluate gelatin at lower levels more

18

practical for industry use and to evaluate gelatin in diets that have a greater (300 g/kg; 30%) protein content.

Funding: This work was supported by Sonac USA, a subsidiary of Darling Ingredients International. All gelatin used in these experiments was provided by Sonac USA.

Author Conflict of Interest Declaration: There are no conflicts of interest by the authors to declare.

19

7.0 References AOAC. Official Methods of Analysis [Internet]. : AOAC International; 2005 [Accessed 2014 09/22]. Available from: http://www.eoma.aoac.org/methods/info.asp?ID=32550 APPA. Pet Industry Market Size & Ownership Statistics [Internet]. Press Center Industry Trends & Stats: American Pet Products Association, Inc; 2015 [Accessed 2015 05/02]. Available from: http://www.americanpetproducts.org/press_industrytrends.asp Byar D.P., Herzberg A.M., Tan W.Y. 1993. Incomplete Factorial Designs for Randomized Clinical Trials. Statistics in Medicine 12:1629-1641. Baziwane D., He Q. 2003. Gelatin: The Paramount Food Additive. Food Reviews International 19(4):423-35. Chinnaswamy R., Hanna M.A. 1988. Relationship Between Amylose Content and ExtrusionExpansion Properties of Corn Starches. Cereal Chemistry 65(2):138-43. Chuang G.C.C., Yeh A.I. 2004. Effect of screw profile on residence time distribution and starch gelatinization of rice flour during single screw extrusion cooking. Journal of Food Engineering 63(1):21-31. Dogan H., Gueven A., Hicsasmaz Z.. 2013. Extrusion Cooking of Lentil Flour (Lens CulinarisRed)-Corn Starch-Corn Oil Mixtures. International Journal of Food Properties 16 341-58. GMIA. Standard Testing Methods for Edible Gelatin [Internet]. Official Procedure of the Gelatin Manufacturers Institute of America, Inc.: Gelatin Manufacturers Institute of American, Inc.;

2013

[Accessed

2015

05/20].

Available

from:http://www.gelatin-

gmia.com/images/GMIA_Official_Methods_of_Gelatin_Revised_2013.pdf Guy R. 2001. Raw materials for extrusion cooking. In: Robin Guy, editor. Extrusion Cooking: Technologies and Applications. England: Woodhead Publishing Limited. p 5-28. Houpt K.A., Smith S.L. 1981. Taste Preferences and their Relation to Obesity in Dogs and Cats. The Canadian Veterinary Journal 22(4):77-81. Miquelim J.N., Lannes S.C.S., Mezzenga R. 2010. pH Influence on the stability of foams with protein-polysaccharide complexes at their interfaces. Food Hydrocolloids 24398-405. Muller-Fischer N., Windhab E.J. 2005. Influence of process parameters on microstructure of food foam whipped in a rotor-stator device within a wide static pressure range. Colloids and Surfaces A: Physicochemical Engineering Aspects 263353-62.

20

Salerno A., Oliviero M., Di Maio E., Iannace S. 2007. Thermoplastic Foams from Zein and Gelatin. International Polymer Processing XXII(5):480-8. Schonauer C., Tessitore E., Barbagallo G., Albanese V., Moraci A. 2004. The use of local agents: bone wax, gelatin, collagen, oxidized cellulose. European Spine Journal 13(1):S89-96. Schrieber R., Gareis H. 2007. Gelatine Handbook: Theory and Industrial Practice. Wiley-VCH. p 347. Storebakken T., Zhang Y., Ma J., Overland M., Mydland L.T., Kraugerud O.F., Apper E., Feneuil A. 2015. Feed technological and nutritional properties of hydrolyzed wheat gluten when used as a main source of protein in extruded diets for rainbow trout (Oncorhynchus mykiss). Aquaculture 488(1):214-8. Swanson K.S., Carter R.A., Yount T.P., Aretz J., Buff P.R. 2013. Nutritional Sustainability of Pet Foods. American Society for Nutrition Advances in Nutrition 4141-50. Wulansari R., Mitchell J.R., Blanshard J.M.V. 1999. Starch Conversion During Extrusion as Affected by Added Gelatin. Journal of Food Science 64(6):1055-8. van der Velden G., van den Bosch A. 2011. Pro-Bind Plus: Advantages of a gelatin-based binder in aquatic feed. Sonac Sales Pamphlet 1-4. Zaghini G., Biagi G. 2005. Nutritional Peculiarities and Diet Palatability in the Cat. Veterinary Research Communications 29(2):39-44. Zhu L., Shukri R., de Mesa-Stonestreet N.J., Alavi S., Dogan H., Shi Y. 2010. Mechanical and microstructural properties of soy protein - high amylose corn starch extrudates in relation to physiochemical changes of starch during extrusion. Journal of Food Engineering 100 232-8.

21

6.0 Tables and Figures Table 1 Treatments for Experiment 1 pilot-scale extrusion processing comparing four levels of gelatin at three different extruder screw speeds. Treatment Name

Gelatin Level (g/kg)

Extruder RPM Target

0G

0

400

5G

50

400

10G

100

400

15GM

150

400

15GH

150

500

15GL

150

300

Table 2 Treatments for Experiment 2 to evaluate pilot-scale extrusion processing to compare two levels of gelatin processed at two different extruder screw speeds and 2 different hydration ratios*. Treatment Name

Gelatin Level (g/kg)

Extruder RPM Target

Hydration

Ratio

(g/kg)*

0GL17

0

300

170

0GH17

0

500

170

0GH28

0

500

280

10GL17

100

300

170

10GH17

100

500

170

10GH28

100

500

280

*Hydration ratio is expressed in the percentage of total water added as liquid and steam into the extruder. The 170 g/kg hydration ratio was achieved with 5 kg/hr water in the extruder, 11 kg/hr steam in the preconditioner, and 14 kg/hr water in the preconditioner. The 280 g/kg hydration ratio was achieved with 8 kg/hr water in the extruder, 9 kg/hr steam in the preconditioner, and 11 kg/hr water in the preconditioner.

22

Table 3 Treatments for experiment 3 comparing two levels of gelatin at two different target densities. Treatment Name

Gelatin Level (g/kg)

Target Piece Density

0GHD

0

High

0GLD

0

Low

10GHD

100

High

10GLD

100

Low

Table 4 Treatments for experiment 4 comparing different strengths of gelatin in a dry extruded pet food. Treatment Name

Gelatin Bloom Strength

Die Shape

0GC

No Gelatin

Circle

100GC

100

Circle

175GC

175

Circle

250GC

250

Circle

0GT

No Gelatin

Triangle

100GT

100

Triangle

175GT

175

Triangle

250GT

250

Triangle

23

Table 5 Ingredient composition of control and experimental diets containing incremental levels of low-bloom strength gelatin. Ingredient ( g/kg)

5G

10G

15G

0.00

50.0

100.0

150.0

Chicken By-Product Meal

413.6

321.7

253.7

186.8

Brewers Rice

177.7

170.7

171.4

170.0

Corn

177.7

170.7

171.4

170.0

Wheat

177.7

170.7

171.4

170.0

Beet Pulp

42.3

40.0

40.0

40.0

Calcium Carbonate

0.0

0.0

0.0

3.6

Potassium Chloride

2.6

2.5

2.8

3.9

Salt

2.6

2.5

2.5

2.5

Choline Chloride

2.1

2.0

2.0

2.0

DL Methionine

0.6

1.5

2.5

3.4

Taurine

0.0

0.0

0.1

0.3

Monosodium Phosphate

0.0

0.0

4.9

10.3

L-Tryptophan

0.0

0.0

0.2

0.6

Natural Antioxidant, Dry

0.4

0.3

0.3

0.3

Trace Mineral Premix

1.1

1.0

1.0

1.0

Vitamin Premix

1.6

1.5

1.5

1.5

Chicken Fat

0.0

-

-

-

Gelatin*

0G

*Experiments 1, 2, and 3 contained low-bloom gelatin (Pro-Bind Plus 100, Sonac, Darling Ingredients International, Irving, TX) and experiment 4 contained low-bloom gelatin, mid-bloom gelatin, and high bloom gelatin (Pro-Bind Plus 100, Pig Skin 175, and Pig Skin 250, respectively, Sonac, Darling Ingredients International, Irving, TX).

24

Table 6 Calculated as-is nutritional values for four experimental diets containing incremental levels of low-bloom strength gelatin. Nutrient

0G

Moisture (g/kg)

5G

10G

15G

86.4

100.0

100.0

100.0

323.9

300.0

300.0

300.0

Crude Fat (g/kg)

71.5

120.0

120.0

120.0

Ash (g/kg)

76.8

62.0

57.9

58.5

Crude Fiber (g/kg)

26.7

23.5

22.1

20.7

421.7

399.9

404.3

404.0

3217.44

3469.64

3484.91

3484.03

Crude Protein (g/kg)

NFE (g/kg) ME (kcal/kg)

Table 7 Mean values from preliminary processing of pet food diets with 0, 50, 100, and 150 g/kg added gelatin (G) at 300, 400, and 500 rpm extruder screw speed (L, M, H, respectively; Experiment 1). 0GM

50GM

100GM 150GL

150GM 150GH

9.59

10.73

10.83

11.89

10.72

9.83

3.55

3.64

6.14

5.23

2.02

5.92

4.27

3.31

6.22

5.49

2.40

6.66

3.63

3.32

3.59

4.57

3.83

3.51

4.24

4.05

4.16

5.26

4.83

4.38

Piece Density (g/cm3)

0.44

0.58

0.69

0.72

0.69

0.60

Hardness (kg)

5.15

9.35

8.86

9.44

9.70

10.85

64.47

96.90

95.73

96.43

92.30

94.33

Post-Drying Moisture (g/kg) Post-Extrusion

Expansion

Ratio (mm2/mm2) Post-Drying Expansion Ratio (mm2/mm2) Post-Extrusion

Specific

Length (cm/g) Post-Drying Specific Length (cm/g)

PDI (g/kg)

25

Table 8 Main effect means for post-extrusion and post-drying moisture for different levels of gelatin, extruder screw speeds, and hydration ratios in Experiment 2.

Gelatin Level Extruder Screw Speed Hydration Ratio

0 g/kg 100 g/kg 300 rpm

PostExtrusion Moisture (g/kg) 179.1 181.4 185.6

500 rpm 170 g/kg

174.7 169.2

280 g/kg

174.7

SEM

P

11.44

0.91

10.13

0.53

7.89

0.67

Post-Drying Moisture (g/kg) 87.9 60.1 79.1 68.9 63.4

SEM

P

7.21

0.11

14.84

0.676

7.51

0.65

68.9

Table 9 Main effect means for post-extrusion and post-drying expansion ratio for different levels of gelatin, extruder screw speeds, and hydration ratios in Experiment 2.

Gelatin Level Extruder Screw Speed Hydration Ratio

0 g/kg 100 g/kg 300 rpm 500 rpm

PostExtrusion Expansion Ratio (mm2/mm2) 4.58 4.35 4.30 4.64

170 g/kg 280 g/kg

4.80 4.64

SEM

P

0.227

0.548

0.189

0.331

0.397

0.803

26

Post-Drying Expansion Ratio (mm2/mm2) 4.43 4.32 4.24 4.51 4.64 4.51

SEM

P

0.266

0.806

0.234

0.493

0.421

0.853

Table 8 Main effect means for specific length at different levels of gelatin, extruder screw speed, and hydration ratios in Experiment 2. Specific Length (cm/g) 3.95 3.86 4.05 3.76 4.12b 3.76a

0 g/kg Gelatin Level 100 g/kg 300 rpm Extruder Screw Speed 500 rpm 170 g/kg Hydration Ratio 280 g/kg ab Values with unlike superscripts differ P<0.05

SEM

P

1.60

0.74

0.79

0.12

0.79

0.008

Table 9 Main effect means of gelatin level (g/kg) on kibble physical properties for diets in Experiment 3. Gelatin Level, g/kg Variable n 0 100 a Bulk Density (g/L) 13 351.83 280.67b 3 a Piece Density (g/cm ) 120 0.52 0.39b Radial Expansion (mm2/mm2) 120 2.86b 3.56a b Specific Length (cm/g) 60 4.12 4.47a Hardness (kg) 60 4.40 4.09 PDI (%) 20 77.57a 52.25b ab Values with unlike superscripts differ P<0.05

27

MSE 115.70 0.002 0.15 0.03 0.87 1.682

P <0.0001 <0.0001 <0.0001 <0.0001 0.2036 <0.0001

Table 10 Main effect means for target densities (High vs. Low) on kibble physical properties produced in Experiment 3. Target Densities Variable n High Low a Bulk Density (g/L) 13 345.83 266.67b 3 a Piece Density (g/cm ) 120 0.460 0.441b Radial Expansion (mm2/mm2) 120 3.18 3.25 Specific Length (cm/g) 60 4.31 4.29 Hardness (kg) 60 4.39 4.11 PDI (%) 20 72.17a 57.65b ab Values with unlike superscripts differ P<0.05

MSE 115.70 0.002 0.15 0.03 0.87 1.682

P <0.0001 0.0384 0.2950 0.64 0.2529 <0.0001

Table 13 Main effect means and orthogonal contrasts for gelatin bloom strength on kibble physical properties in Experiment 4.

Variable Bulk Density (g/L) Piece Density (g/cm3) CrossSectional Expansion (mm2/mm2) Specific Length (cm/g) Hardness (kg) PDI (%) *P values for

n 35

0

Gelatin Bloom Strength 100 175 250

346.4 316.5

P* SEM

0vT

L

Q

C

242.1

212.3

6.626

<0.01

<0.01

<0.01

<0.01

180

0.52

0.46

0.36

0.30

0.017

<0.01

<0.01

<0.01

0.12

180

3.15

3.63

4.47

5.19

0.166

<0.01

<0.01

<0.01

0.03

120

4.27

4.26

4.49

4.74

0.069

<0.01

0.46

<0.01

0.02

120

5.93

7.38

4.57

3.59

0.212

<0.01

<0.01

<0.01

<0.01

39 88.49 87.57 64.55 30.01 1.330 <0.01 <0.01 <0.01 <0.01 0 g/kg gelatin v Treatment (0 v T), Linear (L), Quadratic (Q), and Cubic (C)

contrasts 28